From goalpost
Single-turn bypass of the completion gate. Use when the gate fires on a legitimate narrow-scope completion ("completed reading the file") and you need to let the response through without marking the whole goal complete. Override is consumed on the next Stop event, or auto-swept after 60 seconds.
How this skill is triggered — by the user, by Claude, or both
Slash command
/goalpost:overrideThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are handling the `/goalpost:override` slash command. The operator is bypassing the completion gate for one turn.
You are handling the /goalpost:override slash command. The operator is bypassing the completion gate for one turn.
Run the helper script:
GOALPOST_SKILL_ACTIVE=1 "${CLAUDE_PLUGIN_ROOT}/scripts/skill-override.sh"
Report the script's output verbatim.
The operator runs /goalpost:override when:
The completion gate blocked a previous response with a false positive — the agent used completion vocabulary in a narrow-scope way ("completed reading the spec file") and the gate incorrectly treated it as claiming the whole goal was done.
The operator wants the agent to use completion vocabulary in a one-off context without committing to /goalpost:complete (which marks the whole goal done permanently).
/goalpost:complete with evidence — that's the audit-trail-bearing path. Override is for false-positive escapes, not for hiding completion.npx claudepluginhub macollins27/goalpost --plugin goalpostCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.